Great work and thanks for the inspiration.
In your paper you added just one data point for Gold and concluded that it did fit the model. I figured that if the model is valid also other historic gold data points should fit.
I calculated SF and Market value from 1850–2018 for Gold. The combined Bitcoin Gold data points are shown in the chart below.
As you can see it does not fit the model.
R squared dives from 0,95 to 0,57. As shown in the chart SF for Gold stays more or less the same over a period of a century and a half, but the market value does clearly not.
But Ammous stated, “as money is acquired not for its own purpose, but to be exchanged for other goods and services, its purchasing power is important, not its absolute value”
I recalculated (using official US inflation numbers) all historic market values (Bitcoin and Gold) to reflect its purchasing power, denominated in 2018 USD. For Bitcoin the impact is limited, given its short history. For Gold the impact is massive.
The result is shown in the next chart. The Gold dots are now concentrated around the trendline. R squared is 0,91.
I am not much of a statistician, so I can’t say if with this adjustment the model is still valid. All numbers I used I lookup up the internet, so data integrity is a valid concern.
But with the above points in mind, my conclusion is that Market Value should be swapped for Purchasing Power.
Love to hear your thoughts